Test-time adaptation with semi-supervised learning leverages inference-time homogeneity to maintain AI text detection performance under adversarial humanization, new LLMs, and temporal drift.
The widespread adoption of large language model-assisted writing across society
7 Pith papers cite this work, alongside 3 external citations. Polarity classification is still indexing.
7
Pith papers citing it
3
external citations · external index
citation-role summary
background 3
method 1
citation-polarity summary
representative citing papers
Large-scale audit finds ~9% of US newspaper articles contain AI-generated content, with uneven distribution across outlets and topics and minimal disclosure.
Workshops with over 100 creative writers produced metaphors and four themes for language model governance that favor consent-driven, smaller open models encoding community values.
LLMs extract implicit information more conservatively than humans in social contexts but humans are more conservative in factual contexts, with humans proposing additional triplets overall.
AI writing support reduces ownership most at drafting and least at planning, tracking how much text and ideas the AI contributes, while more AI help raises essay quality.
Interviews reveal a four-stage vibe coding workflow that accelerates prototyping while introducing tensions between quick efficiency and reflective design intention, plus asymmetries in trust and ownership.
Argues for a denoising-first paradigm in LLM-oriented information retrieval, framing challenges via a four-stage progression and providing a taxonomy of signal-to-noise optimization techniques across the pipeline.
citing papers explorer
-
Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift
Test-time adaptation with semi-supervised learning leverages inference-time homogeneity to maintain AI text detection performance under adversarial humanization, new LLMs, and temporal drift.
-
AI use in American newspapers is widespread, uneven, and rarely disclosed
Large-scale audit finds ~9% of US newspaper articles contain AI-generated content, with uneven distribution across outlets and topics and minimal disclosure.
-
Seed Bank, Co-op, Stoop Swap: Metaphors for Governing Language Model Data for Creative Writing
Workshops with over 100 creative writers produced metaphors and four themes for language model governance that favor consent-driven, smaller open models encoding community values.
-
Comparing Human and Large Language Model Interpretation of Implicit Information
LLMs extract implicit information more conservatively than humans in social contexts but humans are more conservative in factual contexts, with humans proposing additional triplets overall.
-
From Planning to Revision: How AI Writing Support at Different Stages Alters Ownership
AI writing support reduces ownership most at drafting and least at planning, tracking how much text and ideas the AI contributes, while more AI help raises essay quality.
-
Vibe Coding in Product Teams: Reconfiguring AI-Assisted Workflows, Prototyping, and Collaboration
Interviews reveal a four-stage vibe coding workflow that accelerates prototyping while introducing tensions between quick efficiency and reflective design intention, plus asymmetries in trust and ownership.
-
LLM-Oriented Information Retrieval: A Denoising-First Perspective
Argues for a denoising-first paradigm in LLM-oriented information retrieval, framing challenges via a four-stage progression and providing a taxonomy of signal-to-noise optimization techniques across the pipeline.